search for: zip_low

...number of special things that > are shuffle like. I'm not sure trying to form a strict hierarchy really works out. The key aspect that breaks this is the use of "undef" in shuffle masks. So there are actually multiple fixed shuffles of the same width which might count as a "zip_lower", depending on the context. I think the distinguishing factor here that means we want to integrate these shuffles into the shufflevector instruction, vs. other sorts of shuffle-ish operations, is that the shuffle pattern is known at compile-time. Given that, optimizations can reason about...

...t;0, 0, 0, 0>) reverse - Reverse the elements of the first operand (<3, 2, 1, 0>) concat - Concatenate the two operands (<0, 1, 2, 3, 4, 5, 6, 7>) split_low - Return the low half of the first operand (<0, 1>) split_high - Return the high half of the first operand (<2, 3>) zip_low - Zip together the low halves of the two operands (<0, 4, 1, 5>) zip_high - Zip together the high halves of the two operands (<2, 6, 3, 7>) unzip_even - Unzip the even elements of the two operands (<0, 2, 4, 6>) unzip_odd - unzip the odd elements of the two operands (<1, 3, 5,...

> -----Original Message----- > From: Chris Lattner <clattner at nondot.org> > Sent: Wednesday, February 5, 2020 4:02 PM > To: Eli Friedman <efriedma at quicinc.com> > Cc: llvm-dev <llvm-dev at lists.llvm.org> > Subject: [EXT] Re: [llvm-dev] [RFC] Extending shufflevector for vscale vectors > (SVE etc.) > > On Jan 29, 2020, at 4:48 PM, Eli Friedman via